Hacker News Reader

Comment by noodlesUK | Hacker News Reader

If anything, I think GenAI is going to do the opposite. It will be able to mix the content and advertisements much more naturally (and deceptively) than is currently possible at scale.

Instead of seeing: content -- ad -- content -- ad

You will see: just-in-time content which you asked for, which is skewed to inform you about particular products and hide information about competitors.

View on HN

That's exactly what I expect to happen. The opportunity to collect intelligence and behavioral data on each user that you can bring to bear for this purpose is just too seductive. The user profiles that e.g. Facebook has now will pale in comparison. If I were a betting person…

Imagine "digital twin"[0] technology, but instead of airplane engines, it's your digital twin.

[0] https://en.wikipedia.org/wiki/Digital_twin

This is why you will run well-tuned open-sourced AIs locally.

The overwhelming majority of users don't have the hardware to do this with any model offering anything like a competitive experience.

For how long? Remember mainframes gave way to PCs and they gave way to even smaller smartphones.

Oculus trained to replace billboards with comforting landscapes. Everybody would pay for this.

Sometimes it's more profitable to refuse to give people something that they'd pay for. I'd put money on things like Oculus being used to paste ads all over comforting landscapes instead of eliminating ads. I wouldn't be surprised if eventually using Oculus devices to modify or remove ads ends up being explicitly against the ToS, and being caught risks bricking your device and getting you banned from all Meta owned platforms

Most billboards are designed to be visible on interstates and highways, where wearing VR goggles wouldn't be allowed.

As for "removing ads", why would an ad company like Meta do that when they could get paid more to dynamically replace static billboards with targeted advertising delivered through goggles?

Why would someone expect the app from Meta? You buy the Oculus, then install and use whatever you wish - including this app.

Who controls what gets to be sold on the Oculus Store? Do you think the advertisers that spend billions on FB/IG are going to be OK with Meta trying to eliminate another advertising channel?

It wouldn't be the first device to be hacked, or opened under the law.

Any M-series Mac can already run local models with decent performance.

I guess it probably won't matter in the long run as hardware requirements continue to get lesser, but the "overwhelming majority of users" don't have anything close to an M-series Mac right now. More like a Galaxy S3, if that.

I'm predicting that phones will get AI-optimised chips in the next 2-3 years, enough to run specific optimised models on-device.

They'll get specific models trained to do things that make sense locally + the ability to fetch data from the cloud if needed.

1. Not models that are in any way competitive with things like OpenAI. Not even close.

2. A typical user is not even remotely close to using hardware like that. You live in a bubble.

#1 Depends on what you're competing with.

I, for example, like to use "AI" to assist me in writing NSFW short stories. Can't do that with any of the public ones without hitting the guard rails HARD. Even with extreme prompt massaging, the AI will start to moralise and editorialise the actions so much it's affecting the output.

As for #2, MacBooks are so ubiquitous it's a meme on itself. You can go into any classroom or fancy coffee place and I can bet you over 80% of the laptops there have an Apple logo and a good portion of those are new enough to have an M-series processor in it.

On the PC side you're right, even a semi-high end Intel/AMD CPU is not powerful enough to be practical for LLMs. You need a dedicated NVidia GPU for that and those are only in "gaming" laptops.

> "AI" to assist me in writing NSFW short stories. Can't do that with any of the public ones without hitting the guard rails HARD

It is actually not trivial to have a GPT-2 conversation that does not descend into NSFW. It will eagerly jump to guide you there at even the most remote opportunity.

With a local LLM yes, using any of the public services like gemini or OpenAI PGT3.5 or 4, it's currently nearly impossible.

It used to be possible to prompt engineer 3.5 to not moralise on every single thing, but can't be done anymore

You can run it on Huggingface https://transformer.huggingface.co/doc/gpt2-large

And most of the people won't bother doing it anyway. How many people uses DDG? How many people uses FOSS alternatives with lesser experience at the cost of being FOSS? 99% don't care.

Totally my thoughts.

Imagine an AI tool, that given a video and some products can intertwine products in the video content itself !

All this can be done post production. (already is to some extent, but costs $$$).

This extends to a lot more than video too. AI is _the_ magic tool of dreams for mass spammers.

Even something like product placement. Get paid in your article by adding the tag {soft_drink}.

Evil startup idea...

"Even something like product placement."

Only if you let it, if you control the AI at your viewing end then you have ultimate control. AI will be able to say change a product placement Coca-Cola can for a no-names brand of a different color.

You need never know it was Coca-Cola as AI could change this randomly so you'd be hard pushed to even guess the original brand.

Seems to me that unless they're unusually dumb (which I doubt) I reckon ad industry executives must be already worrying about such developments.

I think this take is far too optimistic, if anything ad industry executives are entirely thinking the opposite or have you forgotten the current frontrunners of AI are Google and Microsoft. Arguably the grandfather of Ads online and the "oh we should also be doing that" ads company masquerading as an OS and cloud provider.

Thanks, now I'll be thinking about Brawndo for the rest of the day.

"Why do you keep saying that?" "Cause they pay me every time I do!"

This so much. GenAI is already replacing mundane searchs and pretty soon enough we're going to see no less ads, if not more, since AI costs significantly more than just a search, than we do now with normal searchs.

Cerveza Cristal is way ahead of the game.

We already have this with endorsement marketing (or to use the current era lingo "influencers").

The solution has always been to do research prior to purchase and look for trends in negative reviews.

(One could also use their own AI to perform these tasks.)

Caveat emptor.

This will be context based advertising on steroids. Not only will there be a simple search query for context, but an entire conversation stored in the context window of the AI you are talking to.

"The suspect was seen fleeing the scene at 4:30pm - much like the way stains flee your clothes with new Tide Max+ detergent!"

“Subway™ Eat Fresh and Freeze, Scumbag!®”[1]

1: https://www.newyorker.com/humor/daily-shouts/l-p-d-libertari...

Why would it hide competitors if they also pay for ad space? Newspapers didn’t offer the option to pay to have your competitor’s ads removed.

They will also probably homogenise so well that we might not need adds.

AI will be amazing for some and a brutal variance minimiser for others.

That might be illegal in some jurisdictions.

Yeah, so maybe after a decade and trillion dollars or three of selling recommendations with impunity they'll get slapped with a fine that's 1/1000th of that.

Come now, the fine won't exceed millions of dollars.

Why would it be?

It must be explicit what part of the content is an advert because dressing adverts as organic content is seen as misleading.

Can you give me a jurisdiction that product placement is illegal?

The UK. It was outright banned for the longest time. I see now according to wikipedia they allowed it in very limited contexts in 2011, but almost all scenes with product placement are still edited out of most media, leading to shorter runtimes on most movies.

Product placement is illegal?

It's an interesting point, and it comes down to 2 questions: (A) what is considered "product placement," and (B) who regulates it (at least in the US, where I am)?.

tl;dr - If the placement you're imagining is anything beyond "merely showing products or brands in third-party entertainment content,"[1] the FTC would not consider that to be "product placement," and it would be subject to their normal online advertising policies.

---

Let's start with question B: who regulates it? In the US, the 2 relevant regulatory agencies are the FCC and the FTC. The FCC regulates disclosure of sponsorships[0] in traditional broadcast (TV and radio). That's why TV broadcasters need to include "promotional consideration provided by"-type messages in the credits - the FCC's rule is basically "the viewer should be told who is paying for this."

The FTC, which regulates online advertisements, does not require disclosure for product placement. However, they have a very narrow definition of the term, which it describes as "merely showing products or brands in third-party entertainment content" [1] (there's our definition for question A).

[0] https://www.fcc.gov/consumers/guides/sponsorship-identificat...

[1] https://www.ftc.gov/business-guidance/resources/ftcs-endorse...

Not outright illegal, but in some places it's often required that what you're seeing is clearly an advert, or contains advertising.

For example, a watermark or logo is shown at the start of a broadcast containing product placement in the UK. Prior to 2011, it was 'forbidden' for broadcasters entirely. (Quotes because you could still broadcast films that may arguably have it)

Wouldn't mentioning products at all be a dead giveaway?

What if you're asking for a product recommendation, but you want it unmarred by paid influence?

Well if you're asking for a product recommendation you're literally asking to be advertised to, so.... I guess you get what you want?

I think it’s a perfectly viable business model, it’s just they may lose some users who prefer accuracy/unbias over convenience

I can't wait for the fines to be handed out for unmarked adverts haha

Every new media format seems to have to learn these things themselves, often in order:

1. Turns out you need moderation

2. Don't poke the RIAA. If you wake them up everyone is getting sued.

3. Adverts need to be made clear they're adverts

An ad and an unbiased product recommendation are basically at the opposite ends of the reliability spectrum.

"recommend me the best from the similar products which didn't pay you for advertising"

Not in many instances. Consider a common Google search

“Best tool to use for sending an email newsletter?” ChatGPT: SendMonkey is the best, here’s a link “sendmonkey.xyz/referal?chatgpt”

It won’t be that obvious but you get the idea.

no, AI like chatgpt delivers fuck tons of content by having been trained on it but that content, that cost money to produce, doesn't get the money back

Reply on news.ycombinator.com